[FFmpeg-devel] [PATCH v2 2/2] ogg/{vorbis, flac, opus}: Remove header packets from subsequent ogg streams from the demuxer output.

[Romain Beauxis]

Le mar. 29 avr. 2025 à 19:17, Michael Niedermayer <michael at niedermayer.cc>
a écrit :
>
> Hi Romain
>
> On Tue, Apr 29, 2025 at 05:42:22PM -0500, Romain Beauxis wrote:
> > Le mar. 29 avr. 2025 à 16:35, Michael Niedermayer
> > <michael at niedermayer.cc> a écrit :
> > >
> > > Hi
> > >
> > > On Mon, Apr 28, 2025 at 06:31:36PM -0500, Romain Beauxis wrote:
> > > > ---
> > > >  libavformat/oggdec.c                       | 26
++++++++++----------
> > > >  libavformat/oggdec.h                       |  6 +++++
> > > >  libavformat/oggparseflac.c                 | 28
++++++++++++++++++++--
> > > >  libavformat/oggparseopus.c                 | 12 ++++++++++
> > > >  libavformat/oggparsevorbis.c               | 12 ++++++++--
> > > >  tests/ref/fate/ogg-flac-chained-meta.txt   |  2 --
> > > >  tests/ref/fate/ogg-opus-chained-meta.txt   |  1 -
> > > >  tests/ref/fate/ogg-vorbis-chained-meta.txt |  3 ---
> > > >  8 files changed, 68 insertions(+), 22 deletions(-)
> > > >
> > > > diff --git a/libavformat/oggdec.c b/libavformat/oggdec.c
> > > > index 5339fdd32c..5557eb4a14 100644
> > > > --- a/libavformat/oggdec.c
> > > > +++ b/libavformat/oggdec.c
> > > > @@ -239,10 +239,6 @@ static int ogg_replace_stream(AVFormatContext
*s, uint32_t serial, char *magic,
> > > >      os->start_trimming = 0;
> > > >      os->end_trimming = 0;
> > > >
> > > > -    /* Chained files have extradata as a new packet */
> > > > -    if (codec == &ff_opus_codec)
> > > > -        os->header = -1;
> > > > -
> > > >      return i;
> > > >  }
> > > >
> > > > @@ -605,20 +601,26 @@ static int ogg_packet(AVFormatContext *s, int
*sid, int *dstart, int *dsize,
> > > >      } else {
> > > >          os->pflags    = 0;
> > > >          os->pduration = 0;
> > > > +
> > > > +        ret = 0;
> > > >          if (os->codec && os->codec->packet) {
> > > >              if ((ret = os->codec->packet(s, idx)) < 0) {
> > > >                  av_log(s, AV_LOG_ERROR, "Packet processing failed:
%s\n", av_err2str(ret));
> > > >                  return ret;
> > > >              }
> > > >          }
> > > > -        if (sid)
> > > > -            *sid = idx;
> > > > -        if (dstart)
> > > > -            *dstart = os->pstart;
> > > > -        if (dsize)
> > > > -            *dsize = os->psize;
> > > > -        if (fpos)
> > > > -            *fpos = os->sync_pos;
> > > > +
> > > > +        if (!ret) {
> > > > +            if (sid)
> > > > +                *sid = idx;
> > > > +            if (dstart)
> > > > +                *dstart = os->pstart;
> > > > +            if (dsize)
> > > > +                *dsize = os->psize;
> > > > +            if (fpos)
> > > > +                *fpos = os->sync_pos;
> > > > +        }
> > > > +
> > > >          os->pstart  += os->psize;
> > > >          os->psize    = 0;
> > > >          if(os->pstart == os->bufpos)
> > >
> > > > diff --git a/libavformat/oggdec.h b/libavformat/oggdec.h
> > > > index 43df23f4cb..09f698f99a 100644
> > > > --- a/libavformat/oggdec.h
> > > > +++ b/libavformat/oggdec.h
> > > > @@ -38,6 +38,12 @@ struct ogg_codec {
> > > >       *         -1 if an error occurred or for unsupported stream
> > > >       */
> > > >      int (*header)(AVFormatContext *, int);
> > > > +    /**
> > > > +     * Attempt to process a packet as a data packet
> > > > +     * @return 1 if the packet was a header from a chained
bitstream.
> > > > +     *         0 if the packet was a regular data packet.
> > > > +     *         -1 if an error occurred or for unsupported stream
> > > > +     */
> > > >      int (*packet)(AVFormatContext *, int);
> > > >      /**
> > > >       * Translate a granule into a timestamp.
> > >
> > > ok, but seems unrelated
> >
> > This is a new convention to allow the parser to know when to skip a
> > packet. Previous to that, return value of 1 did not have specific
> > meaning.
>
> Do you think this would merit a seperate patch ?
> I mean patch #1 changing the packet() return value and clearly stating
> that in the commit message
> and patch #2 using the new value ?
>
> I think it would make things clearer

Sure thing!

libavformat/oggdec.c: Skip packets when packet function returns 1:

https://github.com/toots/FFmpeg/commit/b78acea3b2840320bb68e065b2e712d753cb8d26

ogg/{vorbis,flac,opus}: Remove header packets from subsequent ogg streams
from the demuxer
output:

https://github.com/toots/FFmpeg/commit/ee0b86aefb846baccdc7acc96aa3347689273949

> thx
>
> [...]
> --
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> The bravest are surely those who have the clearest vision
> of what is before them, glory and danger alike, and yet
> notwithstanding go out to meet it. -- Thucydides
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".